Automatic clipping (#17)

* Early stop when the total time for one track is less than the other track

* Add to demo app

* Fix early stop bug

* Add docs

* Rename AndroiDataSource to DefaultDataSource

* Rename MuxerDataSink to DefaultDataSink

* Add MultiDataSink

Author: natario1

README.md

@@ -131,6 +131,24 @@ Audio: | •••••••••••••••••• source1 ••
 
 And that's all you need to do.
 
+#### Automatic clipping
+
+When concatenating data from multiple sources and on different tracks, it's common to have
+a total audio length that is different than the total video length.
+
+In this case, `Transcoder` will automatically clip the longest track to match the shorter.
+For example:
+
+```java
+Transcoder.into(filePath)
+        .addDataSource(TrackType.VIDEO, video1) // Video, 30 seconds
+        .addDataSource(TrackType.VIDEO, video2) // Video, 30 seconds
+        .addDataSource(TrackType.AUDIO, music) // Audio, 3 minutes
+        // ...
+```
+
+In the situation above, we won't use the full music track, but only the first minute of it.
+
 ## Listening for events
 
 Transcoding will happen on a background thread, but we will send updates through the `TranscoderListener`

demo/src/main/java/com/otaliastudios/transcoder/demo/TranscoderActivity.java

@@ -1,5 +1,6 @@
 package com.otaliastudios.transcoder.demo;
 
+import android.annotation.SuppressLint;
 import android.content.ClipData;
 import android.content.Intent;
 import android.net.Uri;
@@ -13,6 +14,7 @@
 import com.otaliastudios.transcoder.Transcoder;
 import com.otaliastudios.transcoder.TranscoderListener;
 import com.otaliastudios.transcoder.TranscoderOptions;
+import com.otaliastudios.transcoder.engine.TrackType;
 import com.otaliastudios.transcoder.internal.Logger;
 import com.otaliastudios.transcoder.strategy.DefaultAudioStrategy;
 import com.otaliastudios.transcoder.strategy.DefaultVideoStrategy;
@@ -39,6 +41,7 @@ public class TranscoderActivity extends AppCompatActivity implements
 
     private static final String FILE_PROVIDER_AUTHORITY = "com.otaliastudios.transcoder.demo.fileprovider";
     private static final int REQUEST_CODE_PICK = 1;
+    private static final int REQUEST_CODE_PICK_AUDIO = 5;
     private static final int PROGRESS_BAR_MAX = 1000;
 
     private RadioGroup mAudioChannelsGroup;
@@ -48,20 +51,24 @@ public class TranscoderActivity extends AppCompatActivity implements
     private RadioGroup mVideoAspectGroup;
     private RadioGroup mVideoRotationGroup;
     private RadioGroup mSpeedGroup;
+    private RadioGroup mAudioReplaceGroup;
 
     private ProgressBar mProgressView;
     private TextView mButtonView;
+    private TextView mAudioReplaceView;
 
     private boolean mIsTranscoding;
     private Future<Void> mTranscodeFuture;
     private Uri mTranscodeInputUri1;
     private Uri mTranscodeInputUri2;
     private Uri mTranscodeInputUri3;
+    private Uri mAudioReplacementUri;
     private File mTranscodeOutputFile;
     private long mTranscodeStartTime;
     private TrackStrategy mTranscodeVideoStrategy;
     private TrackStrategy mTranscodeAudioStrategy;
 
+    @SuppressLint("SetTextI18n")
     @Override
     protected void onCreate(Bundle savedInstanceState) {
         super.onCreate(savedInstanceState);
@@ -83,20 +90,32 @@ protected void onCreate(Bundle savedInstanceState) {
         mProgressView = findViewById(R.id.progress);
         mProgressView.setMax(PROGRESS_BAR_MAX);
 
+        mAudioReplaceView = findViewById(R.id.replace_info);
+
         mAudioChannelsGroup = findViewById(R.id.channels);
         mVideoFramesGroup = findViewById(R.id.frames);
         mVideoResolutionGroup = findViewById(R.id.resolution);
         mVideoAspectGroup = findViewById(R.id.aspect);
         mVideoRotationGroup = findViewById(R.id.rotation);
         mSpeedGroup = findViewById(R.id.speed);
         mAudioSampleRateGroup = findViewById(R.id.sampleRate);
+        mAudioReplaceGroup = findViewById(R.id.replace);
 
         mAudioChannelsGroup.setOnCheckedChangeListener(this);
         mVideoFramesGroup.setOnCheckedChangeListener(this);
         mVideoResolutionGroup.setOnCheckedChangeListener(this);
         mVideoAspectGroup.setOnCheckedChangeListener(this);
         mAudioSampleRateGroup.setOnCheckedChangeListener(this);
         syncParameters();
+
+        mAudioReplaceGroup.setOnCheckedChangeListener((group, checkedId) -> {
+            mAudioReplacementUri = null;
+            mAudioReplaceView.setText("No replacement selected.");
+            if (checkedId == R.id.replace_yes && !mIsTranscoding) {
+                startActivityForResult(new Intent(Intent.ACTION_GET_CONTENT)
+                        .setType("audio/*"), REQUEST_CODE_PICK_AUDIO);
+            }
+        });
     }
 
     @Override
@@ -173,6 +192,14 @@ protected void onActivityResult(int requestCode, int resultCode, final Intent da
                 transcode();
             }
         }
+        if (requestCode == REQUEST_CODE_PICK_AUDIO
+                && resultCode == RESULT_OK
+                && data != null
+                && data.getData() != null) {
+            mAudioReplacementUri = data.getData();
+            mAudioReplaceView.setText(mAudioReplacementUri.toString());
+
+        }
     }
 
     private void transcode() {
@@ -208,9 +235,16 @@ private void transcode() {
         mTranscodeStartTime = SystemClock.uptimeMillis();
         setIsTranscoding(true);
         TranscoderOptions.Builder builder = Transcoder.into(mTranscodeOutputFile.getAbsolutePath());
-        if (mTranscodeInputUri1 != null) builder.addDataSource(this, mTranscodeInputUri1);
-        if (mTranscodeInputUri2 != null) builder.addDataSource(this, mTranscodeInputUri2);
-        if (mTranscodeInputUri3 != null) builder.addDataSource(this, mTranscodeInputUri3);
+        if (mAudioReplacementUri == null) {
+            if (mTranscodeInputUri1 != null) builder.addDataSource(this, mTranscodeInputUri1);
+            if (mTranscodeInputUri2 != null) builder.addDataSource(this, mTranscodeInputUri2);
+            if (mTranscodeInputUri3 != null) builder.addDataSource(this, mTranscodeInputUri3);
+        } else {
+            if (mTranscodeInputUri1 != null) builder.addDataSource(TrackType.VIDEO, this, mTranscodeInputUri1);
+            if (mTranscodeInputUri2 != null) builder.addDataSource(TrackType.VIDEO, this, mTranscodeInputUri2);
+            if (mTranscodeInputUri3 != null) builder.addDataSource(TrackType.VIDEO, this, mTranscodeInputUri3);
+            builder.addDataSource(TrackType.AUDIO, this, mAudioReplacementUri);
+        }
         mTranscodeFuture = builder.setListener(this)
                 .setAudioTrackStrategy(mTranscodeAudioStrategy)
                 .setVideoTrackStrategy(mTranscodeVideoStrategy)

demo/src/main/res/layout/activity_transcoder.xml

@@ -285,12 +285,51 @@
                     android:layout_height="wrap_content" />
             </RadioGroup>
 
+            <!-- REPLACE AUDIO -->
+            <TextView
+                android:padding="16dp"
+                android:layout_width="match_parent"
+                android:layout_height="wrap_content"
+                android:text="Replace audio" />
+            <RadioGroup
+                android:id="@+id/replace"
+                android:checkedButton="@id/replace_no"
+                android:orientation="horizontal"
+                android:gravity="center"
+                android:layout_width="match_parent"
+                android:layout_height="wrap_content">
+                <com.google.android.material.radiobutton.MaterialRadioButton
+                    android:id="@+id/replace_no"
+                    android:text="No"
+                    android:paddingLeft="8dp"
+                    android:paddingRight="8dp"
+                    android:layout_width="wrap_content"
+                    android:layout_height="wrap_content" />
+                <com.google.android.material.radiobutton.MaterialRadioButton
+                    android:id="@+id/replace_yes"
+                    android:text="Yes (choose source)"
+                    android:paddingLeft="8dp"
+                    android:paddingRight="8dp"
+                    android:layout_width="wrap_content"
+                    android:layout_height="wrap_content" />
+            </RadioGroup>
+            <TextView
+                android:id="@+id/replace_info"
+                android:text="No replacement selected."
+                android:maxLines="1"
+                android:ellipsize="end"
+                android:paddingLeft="32dp"
+                android:paddingRight="32dp"
+                android:layout_gravity="center"
+                android:layout_width="wrap_content"
+                android:layout_height="wrap_content" />
+
             <!-- INFO TEXT -->
             <TextView
                 android:padding="16dp"
                 android:layout_width="match_parent"
                 android:layout_height="wrap_content"
-                android:text="Note: our API offers many more options than these!" />
+                android:text="Note: our API offers many more options than these!\nYou can select more than one video, in which case they will be concatenated together." />
 
             <!-- SPACE AND BUTTONS -->
             <Space

lib/src/main/java/com/otaliastudios/transcoder/engine/Engine.java

@@ -22,7 +22,7 @@
 import com.otaliastudios.transcoder.internal.ValidatorException;
 import com.otaliastudios.transcoder.sink.DataSink;
 import com.otaliastudios.transcoder.sink.InvalidOutputFormatException;
-import com.otaliastudios.transcoder.sink.MediaMuxerDataSink;
+import com.otaliastudios.transcoder.sink.DefaultDataSink;
 import com.otaliastudios.transcoder.source.DataSource;
 import com.otaliastudios.transcoder.strategy.TrackStrategy;
 import com.otaliastudios.transcoder.time.TimeInterpolator;
@@ -37,7 +37,6 @@
 import androidx.annotation.Nullable;
 
 import java.util.ArrayList;
-import java.util.Arrays;
 import java.util.HashSet;
 import java.util.List;
 import java.util.Set;
@@ -205,7 +204,6 @@ private void closeCurrentStep(@NonNull TrackType type) {
     private TrackTranscoder getCurrentTrackTranscoder(@NonNull TrackType type, @NonNull TranscoderOptions options) {
         int current = mCurrentStep.require(type);
         int last = mTranscoders.require(type).size() - 1;
-        int max = mDataSources.require(type).size();
         if (last == current) {
             // We have already created a transcoder for this step.
             // But this step might be completed and we might need to create a new one.
@@ -251,26 +249,49 @@ public long interpolate(@NonNull TrackType type, long time) {
         };
     }
 
-    private double getTrackProgress(@NonNull TrackType type) {
-        if (!mStatuses.require(type).isTranscoding()) return 0.0D;
+    private long getTrackDurationUs(@NonNull TrackType type) {
+        if (!mStatuses.require(type).isTranscoding()) return 0L;
         int current = mCurrentStep.require(type);
         long totalDurationUs = 0;
-        long completedDurationUs = 0;
         for (int i = 0; i < mDataSources.require(type).size(); i++) {
             DataSource source = mDataSources.require(type).get(i);
-            if (i < current) {
+            if (i < current) { // getReadUs() is a better approximation for sure.
                 totalDurationUs += source.getReadUs();
-                completedDurationUs += source.getReadUs();
-            } else if (i == current) {
-                totalDurationUs += source.getDurationUs();
-                completedDurationUs += source.getReadUs();
             } else {
                 totalDurationUs += source.getDurationUs();
-                completedDurationUs += 0;
             }
         }
-        if (totalDurationUs == 0) totalDurationUs = 1;
-        return (double) completedDurationUs / (double) totalDurationUs;
+        return totalDurationUs;
+    }
+
+    private long getTotalDurationUs() {
+        boolean hasVideo = hasVideoSources() && mStatuses.requireVideo().isTranscoding();
+        boolean hasAudio = hasAudioSources() && mStatuses.requireVideo().isTranscoding();
+        long video = hasVideo ? getTrackDurationUs(TrackType.VIDEO) : Long.MAX_VALUE;
+        long audio = hasAudio ? getTrackDurationUs(TrackType.AUDIO) : Long.MAX_VALUE;
+        return Math.min(video, audio);
+    }
+
+    private long getTrackReadUs(@NonNull TrackType type) {
+        if (!mStatuses.require(type).isTranscoding()) return 0L;
+        int current = mCurrentStep.require(type);
+        long completedDurationUs = 0;
+        for (int i = 0; i < mDataSources.require(type).size(); i++) {
+            DataSource source = mDataSources.require(type).get(i);
+            if (i <= current) {
+                completedDurationUs += source.getReadUs();
+            }
+        }
+        return completedDurationUs;
+    }
+
+    private double getTrackProgress(@NonNull TrackType type) {
+        if (!mStatuses.require(type).isTranscoding()) return 0.0D;
+        long readUs = getTrackReadUs(type);
+        long totalUs = getTotalDurationUs();
+        LOG.v("getTrackProgress - readUs:" + readUs + ", totalUs:" + totalUs);
+        if (totalUs == 0) totalUs = 1; // Avoid NaN
+        return (double) readUs / (double) totalUs;
     }
 
     /**
@@ -281,7 +302,7 @@ private double getTrackProgress(@NonNull TrackType type) {
      * @throws InterruptedException when cancel to transcode
      */
     public void transcode(@NonNull TranscoderOptions options) throws InterruptedException {
-        mDataSink = new MediaMuxerDataSink(options.getOutputPath());
+        mDataSink = new DefaultDataSink(options.getOutputPath());
         mDataSources.setVideo(options.getVideoDataSources());
         mDataSources.setAudio(options.getAudioDataSources());
 
@@ -295,16 +316,7 @@ public void transcode(@NonNull TranscoderOptions options) throws InterruptedExce
             }
         }
 
-        // Compute total duration: it is the minimum between the two.
-        long audioDurationUs = hasAudioSources() ? 0 : Long.MAX_VALUE;
-        long videoDurationUs = hasVideoSources() ? 0 : Long.MAX_VALUE;
-        for (DataSource source : options.getVideoDataSources()) videoDurationUs += source.getDurationUs();
-        for (DataSource source : options.getAudioDataSources()) audioDurationUs += source.getDurationUs();
-        long totalDurationUs = Math.min(audioDurationUs, videoDurationUs);
-        LOG.v("Duration (us): " + totalDurationUs);
-
-        // TODO if audio and video have different lengths, we should clip the longer one!
-        // TODO ClipDataSource or something like that, to choose
+        // TODO ClipDataSource or something like that
 
         // Compute the TrackStatus.
         int activeTracks = 0;
@@ -314,6 +326,7 @@ public void transcode(@NonNull TranscoderOptions options) throws InterruptedExce
         TrackStatus audioStatus = mStatuses.requireAudio();
         if (videoStatus.isTranscoding()) activeTracks++;
         if (audioStatus.isTranscoding()) activeTracks++;
+        LOG.v("Duration (us): " + getTotalDurationUs());
 
         // Pass to Validator.
         //noinspection UnusedAssignment
@@ -331,22 +344,35 @@ public void transcode(@NonNull TranscoderOptions options) throws InterruptedExce
             long loopCount = 0;
             boolean stepped = false;
             boolean audioCompleted = false, videoCompleted = false;
+            boolean forceAudioEos = false, forceVideoEos = false;
+            double audioProgress = 0, videoProgress = 0;
             while (!(audioCompleted && videoCompleted)) {
                 if (Thread.interrupted()) {
                     throw new InterruptedException();
                 }
                 stepped = false;
+
+                // First, check if we have to force an input end of stream for some track.
+                // This can happen, for example, if user adds 1 minute (video only) with 20 seconds
+                // of audio. The video track must be stopped once the audio stops.
+                long totalUs = getTotalDurationUs() + 100 /* tolerance */;
+                forceAudioEos = getTrackReadUs(TrackType.AUDIO) > totalUs;
+                forceVideoEos = getTrackReadUs(TrackType.VIDEO) > totalUs;
+
+                // Now step for transcoders that are not completed.
                 audioCompleted = isCompleted(TrackType.AUDIO);
                 videoCompleted = isCompleted(TrackType.VIDEO);
                 if (!audioCompleted) {
-                    stepped |= getCurrentTrackTranscoder(TrackType.AUDIO, options).transcode();
+                    stepped |= getCurrentTrackTranscoder(TrackType.AUDIO, options).transcode(forceAudioEos);
                 }
                 if (!videoCompleted) {
-                    stepped |= getCurrentTrackTranscoder(TrackType.VIDEO, options).transcode();
+                    stepped |= getCurrentTrackTranscoder(TrackType.VIDEO, options).transcode(forceVideoEos);
                 }
                 if (++loopCount % PROGRESS_INTERVAL_STEPS == 0) {
-                    setProgress((getTrackProgress(TrackType.VIDEO)
-                            + getTrackProgress(TrackType.AUDIO)) / activeTracks);
+                    audioProgress = getTrackProgress(TrackType.AUDIO);
+                    videoProgress = getTrackProgress(TrackType.VIDEO);
+                    LOG.v("progress - video:" + videoProgress + " audio:" + audioProgress);
+                    setProgress((videoProgress + audioProgress) / activeTracks);
                 }
                 if (!stepped) {
                     Thread.sleep(SLEEP_TO_WAIT_TRACK_TRANSCODERS);

lib/src/main/java/com/otaliastudios/transcoder/sink/DataSink.java

@@ -13,7 +13,7 @@
 
 /**
  * A DataSink is an abstract representation of an encoded data collector.
- * Currently the only implementation is {@link MediaMuxerDataSink} which collects
+ * Currently the only implementation is {@link DefaultDataSink} which collects
  * data into a {@link java.io.File} using {@link android.media.MediaMuxer}.
  *
  * However there might be other implementations in the future, for example to stream data